Full-Stack Software Engineer

Responsibilities

• Building and extending the offline, standalone application (Python/Flask) that ingests documents, images, audio, and video and runs locally hosted AI models to translate, summarize, extract entities, and catalog content
• Developing the browser-based local user interface ingest dashboard, tag filters, search, and a natural-language question interface
• Integrating and operating on-device AI/ML components, including local open-weight language models, retrieval-augmented generation, a vector store and embeddings, speech transcription, and multimodal vision models
• Building the controlled sandbox and data-security features malware scanning and rule-based detection, file-type allow-listing, a quarantine landing zone, unprivileged containers with read-only mounts, and ephemeral teardown on shutdown
• Packaging the application for standalone, offline installation on a zero-trust Linux laptop baseline, including single-package and side-load delivery across enclave boundaries
• Maintaining a self-hosted CI/CD pipeline and cutting versioned releases that record model and behavior changes
• Validating near-real-time processing against edge data loads from 128 GB to 1 TB across text, image, audio, and video
• Supporting redeployment and iteration of the capability after the prototype as requirements change

Requirements

• Front-end web development with JavaScript, HTML, and CSS for browser-based interfaces
• Strong Python, including backend web development (Flask or equivalent)
• Containerization with Docker and comfort working on Linux (Rocky Linux or the RHEL family)
• Experience integrating AI/ML or large language model components into applications
• Experience with data workflows using relational databases and/or vector stores

Qualifications

• Local or offline LLM tooling (e.g., Ollama) and retrieval-augmented generation (RAG) Vector databases (e.g., ChromaDB) and sentence embeddings (e.g., all-MiniLM) Speech-to-text (Whisper / faster-whisper) and multimodal vision models Text extraction (Apache Tika) and metadata standards (Dublin Core) Application security and sandboxing ClamAV, YARA, unprivileged containers, encryption, role-based access control Self-hosted CI/CD and Git for source control Building for air-gapped, disconnected, or edge environments and standalone portable devices

Benefits

• Health/Dental/Vision
• 401(k)
• Paid Time Off
• STD/LTD/Life Insurance/Voluntary Life Insurance
• Stipends
• Referral Bonuses